Male-initiated sexual aggression toward female acquaintances is a major public-health problem in adolescence and early adulthood, and risk of sexual aggression is strongly associated with alcohol use. The efficacy of existing prevention programs for sexual aggression has been disappointing, suggesting that additional pathways to change should be investigated. Because alcohol intoxication is such an important contextual variable for sexual aggression, it is also important to evaluate the generalizability of any changes to the intoxicated state. The current project seeks to translate emerging methods from cognitive and learning sciences into a novel cognitive-training procedure and to evaluate the degree to which training effects are maintained during intoxication in the laboratory. Ultimately, we hope that extended versions of the proposed training procedure could be used to supplement existing prevention efforts in sexual aggression. The goals are (1) To evaluate whether a novel cognitive-training procedure enhances men's learning about women's social cues and whether those training effects transfer to other tasks assessing attention, decision-making, and behavioral intention to exhibit sexual aggression; (2) To examine alcohol effects on men's processing of women's affective cues, their decision making, and their behavioral intention to exhibit sexual aggression; and (3) To determine whether the training program moderates or reduces alcohol effects on attention, decision making and behavioral intention to exhibit sexual aggression. One-hundred eighty 21- 25 year old participants will complete either the training program or a control task, after which they will be randomly assigned to consume either no alcohol or a moderate dose of alcohol. Both groups will complete social-cognitive tasks assessing attention, decision making, and behavioral intention. In another session, participants will complete assessments of drinking patterns, alcohol expectancies, rape supportive attitudes, and past history of sexual aggression. Multilevel and general linear modeling will be used to evaluate the influence of training condition and intoxication on cognitive processing as well as the interactions of training and alcohol effects. The explicit translational goal of the project isto take emerging theory and methods from basic cognitive and learning science and develop them into procedures with the potential of contributing to effective prevention efforts in sexual aggression, and which will be robust to the important contextual effects of alcohol intoxication.